U.S. patent number 5,878,160 [Application Number 08/589,102] was granted by the patent office on 1999-03-02 for flow type particle image analyzing method and apparatus for displaying particle images by classifying them based on their configurational features.
This patent grant is currently assigned to Hitachi, Ltd.. Invention is credited to Hideki Asai, Hideyuki Horiuchi, Yasuaki Kojima, Norio Oowada, Kazuhiro Sano.
United States Patent |
5,878,160 |
Horiuchi , et al. |
March 2, 1999 |
Flow type particle image analyzing method and apparatus for
displaying particle images by classifying them based on their
configurational features
Abstract
A flow type particle image analytical method for feeding a
particles-suspended sample by surrounding it by a cleaning
solution, imaging particles in the sample by irradiating light to
it, and automatically classifying the particles in the sample by
analyzing the picked-up particle images, wherein the method
comprises a means for designating the kind of particles to be
reviewed beforehand among the classified particles, a means for
storing only particles which are automatically classified to the
designated particle kind in a review image memory, a means for
displaying particles of the same kind on a CRT display in batch so
as to review particle images, and a means for classifying each of
the displayed particles finally by an operator or for changing the
particle name.
Inventors: |
Horiuchi; Hideyuki (Abiko,
JP), Asai; Hideki (Mito, JP), Kojima;
Yasuaki (Hitachinaka, JP), Oowada; Norio
(Hitachinaka, JP), Sano; Kazuhiro (Naka-machi,
JP) |
Assignee: |
Hitachi, Ltd. (Tokyo,
JP)
|
Family
ID: |
11874142 |
Appl.
No.: |
08/589,102 |
Filed: |
January 23, 1996 |
Foreign Application Priority Data
|
|
|
|
|
Feb 1, 1995 [JP] |
|
|
7-014910 |
|
Current U.S.
Class: |
382/133; 382/224;
702/21 |
Current CPC
Class: |
G01N
15/1459 (20130101); G01N 15/1463 (20130101) |
Current International
Class: |
G01N
15/14 (20060101); G06K 009/78 () |
Field of
Search: |
;382/133,134,128,224,225,203,309,311 ;364/555 ;702/21 |
References Cited
[Referenced By]
U.S. Patent Documents
Foreign Patent Documents
|
|
|
|
|
|
|
A0549 905 |
|
Jan 1992 |
|
EP |
|
63-94156 |
|
Apr 1988 |
|
JP |
|
3-105235 |
|
May 1991 |
|
JP |
|
4-72544 |
|
Mar 1992 |
|
JP |
|
4-309841 |
|
Nov 1992 |
|
JP |
|
7-20124 |
|
Jan 1995 |
|
JP |
|
WO91/15826 |
|
Oct 1991 |
|
WO |
|
Primary Examiner: Chang; Jon
Attorney, Agent or Firm: Fay, Sharpe, Beall, Fagan, Minnich
& McKee
Claims
What is claimed is:
1. A flow type particle image analytical apparatus having a forming
means for forming a sample flow by flowing a sample liquid
containing particles, a generating means for generating images of
said particles in said sample liquid by irradiating light to said
sample flow, and an automatic classifying means for classifying
information relating to said particles obtained by analyzing said
images detected based on plural configurational features relating
to configuration, size or color of the image, wherein said
apparatus further comprises
a selecting means for selecting at least one to be reviewed of the
particles, as to which information has been classified,
a review image memory for storing an image of said particles
corresponding to said selecting,
a display means for displaying said image stored in said review
image memory for reviewing said image, whereby said information
relating to said particles which are selected and are not selected
to be reviewed is corrected based on said image stored in said
review image memory,
a registering means for registering a specific particle kind
beforehand, and
an image storage control means including an image storage ON/OFF
means for setting ON/OFF of storage of an image of said particles
corresponding to said specific particle kind and storing said image
of said particles which correspond to said specific particle kind
and are automatically classified, in said review image memory when
said image storage ON/OFF means is in the ON state.
2. A flow type particle image analytical apparatus according to
claim 1, which further comprises
an image storage control means including a means for setting ON/OFF
of storage of an image of ambiguous particles which are considered
to correspond to a particle kind of a first candidate but are
considered to possibly correspond to a particle kind of a second
candidate close to said particle kind of said first candidate and
for storing said image of said ambiguous particles in said review
image memory, when said image storage ON/OFF means is in the ON
state, and wherein
said display means is structured so as to display said kinds of
both said candidates in addition to said image of ambiguous
particles which is stored as an image relating to said particle
kinds of both said candidates.
3. A flow type particle image analytical apparatus according to
claim 1, which further comprises
an image storage ON/OFF means for setting ON/OFF of storage of an
image of said particles and storing an image of particles
corresponding to another particle kind which is apt to be
misdecided as one corresponding to said designated particle kind by
automatic classification in said review image memory when said
image storage ON/OFF means is in the ON state, and wherein
said display means is structured so as to display said image of
particles corresponding to said designated particle kind and said
image of particles corresponding to said another particle kind for
comparison observation.
4. A flow type particle image analytical apparatus according to
claim 1, wherein said
particles are displayed on the display unit and finally classified
or the particle kind name thereof is changed and furthermore said
automatic classification result of particles which are
automatically classified is corrected on the basis of said
classification result.
5. A flow type particle image analytical apparatus according to
claim 4, wherein said automatic classification result of said
automatically classified particles is corrected on the basis of the
existence ratio of said particles for each corresponding particle
kind.
6. A flow type particle image analytical apparatus according to
claim 1, wherein said particles are cells of an organism.
7. A flow type particle image analytical apparatus according to
claim 1, wherein said particles are blood corpuscles in blood.
8. A flow type particle image analytical apparatus according to
claim 1, wherein said particles are sediment of urine.
Description
BACKGROUND OF THE INVENTION
The present invention relates to a flow type particle image
analytical method and apparatus for generating an image of
particles in a sample, which form a suspended flow of the sample,
by irradiating light to the formed sample flow and automatically
classifying the particles in the sample by analyzing the generated
particle image, and particularly to a flow type particle image
analytical method and apparatus suited to classify and display the
particle image based on configurational features such as
configuration, size or color of the particle image according to
biological kinds of particles such as cells in blood or urine.
To conventionally classify and analyze particles such as cells in
blood or cells existing in urine, a sample is prepared on a glass
slide sheet and observed with a microscope. In the case of urine,
the particle concentration in the urine is thin, so that a
measuring sample is centrifugalized and concentrated by a
centrifuge beforehand and then observed. In an apparatus for
automating these observation and examination work, a sample of
blood or other material is coated on a glass slide and set in a
microscope, and the microscope stage is automatically scanned and
stopped at a position where particles exist. A still image of
particles then is picked up and particles in the sample are
classified by using a characteristic extraction and pattern
recognition method by the image processing art.
However, in the aforementioned method, it takes a lot of time to
prepare a sample and an operation for finding a particle by moving
the microscope stage mechanically and moving the particle into a
suitable image fetching area is required. As a result, problems
arise in that the analysis requires a lot of time and the mechanism
is complicated.
To realize high precise examination and labor saving, there is a
flow type particle image analytical apparatus using a flow cell for
using a sheath solution which has a cleaning solution as an outer
layer and controlling the sample solution to an extremely flat flow
which is disclosed in, for example, U.S. Pat. No. 4,338,024,
Japanese Patent Application Laid-Open 63-94156, and Japanese Patent
Application Laid-Open 4-72544.
In the flow type particle image analytical apparatus, a sample
moving in a flow cell is imaged, for example, by a video camera and
the particles in the sample are classified and counted by
processing the picked-up still image.
As a flow type particle image analytical apparatus for imaging
particles in a sample by changing the magnification, a particle
analytical apparatus is described in Japanese Patent Application
Laid-Open 3-105235 and Japanese Patent Application Laid-Open
4-309841.
The flow type particle analytical apparatus described in Japanese
Patent Application Laid-Open 3-105235 and Japanese Patent
Application Laid-Open 4-309841 mentioned above comprises a
stroboscope continuously emitting light having a short emission
time, a stop for adjusting the quantity of the stroboscopic light,
a diffusion plate for eliminating variations in the quantity of
stroboscopic light, a condenser lens for focusing the stroboscopic
light, a flow cell which is arranged in the position where the
stroboscopic light passes and lets a sample flow flat by
surrounding the sample by a sheath solution, an object lens for
forming an image of a particle irradiated by the stroboscopic
light, a high-powered projection lens, a low-powered projection
lens, a TV camera for picking up an image, a means for moving the
diffusion plate, a means for changing the aperture stop, and a
switching means for switching the high-powered projection lens and
low-powered projection lens.
A method for classifying a particle image formed by the
aforementioned flow type particle image analytical apparatus by the
particle size and displaying it on a CRT screen and classifying
particles by an operator is proposed in U.S. Pat. No.
4,612,614.
However, there is the following problem imposed in the flow type
particle image analytical apparatus described in U.S. Pat. No.
4,612,614 mentioned above.
Namely, it is premised in the flow type particle image analytical
apparatus that the particle classification work is reviewed by an
operator by watching an image on the CRT and an extremely large
image memory is necessary so as to analyze all particles in a
sample. Particularly in a sample having a large particle
concentration, all particles cannot be stored in a limited image
memory and furthermore it requires a lot of time for an operator to
proceed with the classification work by watching images. When the
image memory is not sufficient, the number of particles to be
reviewed decreases and it is difficult to expect a reproducible
classification result.
SUMMARY OF THE INVENTION
An object of the present invention is to provide a flow type
particle image analytical method and apparatus suited to
realization of a small-capacity image memory and prevention of
reduction in the reproducibility of the classification result.
The problem solving means so as to accomplish the object of the
present invention are as follows:
1. A flow type particle image analytical method characterized in
that the method forms a flow of a particles-suspended sample,
generates an image of particles in the sample by irradiating light
to the formed sample flow, and automatically classifies the
particles in the sample by analyzing the generated particle image,
wherein the method designates the kind of particles to be reviewed,
controls storage of an image in a review image memory so as to
store an image of particles corresponding to the designated
particle kind in the review image memory, and displays the stored
particle image on the display for review (Claim 1).
2. A flow type particle image analytical method of the solution
means of 1 characterized in that the aforementioned particle image
display is given in batch for each particle kind (Claim 2).
3. A flow type particle image analytical method of the solution
means of 1 characterized in that the aforementioned particle kind
designation is made for one sample in each measuring mode (Claim
3).
4. A flow type particle image analytical apparatus characterized in
that the apparatus forms a flow of a particles-suspended sample,
generates an image of particles in the sample by irradiating light
to the formed sample flow, and automatically classifies the
particles in the sample by analyzing the generated particle image,
wherein the apparatus comprises a means for designating the kind of
particles to be reviewed, a review image memory, a means for
controlling storage of an image in the review image memory so as to
store an image of particles corresponding to the designated
particle kind in the review image memory, and a means for
displaying the stored particle image for review (Claim 4).
5. A flow type particle image analytical apparatus of the solution
means of 4 characterized in that the aforementioned image storage
control means includes a means for turning storage of an image of
particles which cannot be classified on or off and is structured so
as to store the image of particles which cannot be classified in
the review image memory when the image storage ON/OFF means is in
the ON state (Claim 5).
6. A flow type particle image analytical apparatus of the solution
means of 4 characterized in that the apparatus has a means for
registering a specific particle kind beforehand and the
aforementioned image storage control means includes a means for
turning storage of an image of particles corresponding to the
specific particle kind on or off and is structured so as to store
an image of particles which correspond to the registered specific
particle kind and are automatically classified in the review image
memory when the image storage ON/OFF means is in the ON state
(Claim 6).
7. A flow type particle image analytical apparatus of the solution
means of 4 characterized in that the aforementioned image storage
control means includes a means for turning storage of an image of
ambiguous particles which are considered to correspond to the
particle kind of the first candidate but are considered to possibly
correspond to the particle kind of the second candidate close to
the particle kind of the first candidate on or off and is
structured so as to store the image of ambiguous particles in the
review image memory when the image storage ON/OFF means is in the
ON state and the aforementioned display means is structured so as
to display the particle kinds of both the candidates in addition to
the image of ambiguous particles which is stored as an image
relating to the particle kinds of both the candidates (Claim
7).
8. A flow type particle image analytical apparatus of the solution
means of 4 characterized in that the aforementioned image storage
control means includes a means for turning storage of an image of
particles on or off and is structured so as to store an image of
particles corresponding to another particle kind which is apt to be
misdecided as one corresponding to the designated particle kind by
automatic classification in the review image memory when the image
storage ON/OFF means is in the ON state and the aforementioned
display means is structured so as to display the image of particles
corresponding to the designated particle kind and the image of
particles corresponding to the another particle kind for comparison
observation (Claim 8).
9. A flow type particle image analytical apparatus of the solution
means of 4 characterized in that the apparatus is structured so
that an operator classifies particles displayed on the display unit
finally or changes the particle kind name and furthermore corrects
the automatic classification result of particles which are
automatically classified on the basis of the aforementioned
classification result (Claim 9).
10. A flow type particle image analytical apparatus of the solution
means of 9 characterized in that the apparatus is structured so
that the automatic classification result of the automatically
classified particles is corrected on the basis of the existence
ratio of the particles which are objects of review for each
corresponding particle kind (Claim 10).
11. A flow type particle image analytical apparatus of the solution
means of 4 characterized in that the particles are cells of an
organism (Claim 11).
12. A flow type particle image analytical apparatus of the solution
means of 4 characterized in that the particles are blood corpuscles
in blood (Claim 12).
13. A flow type particle image analytical apparatus of the solution
means of 4 characterized in that the particles are sediment of
urine (Claim 13).
The solution means of 1 designates the biological kind of particles
to be reviewed, controls storage of an image in a review image
memory so as to store an image of particles corresponding to the
designated particle kind in the review image memory, and displays
the stored particle image on the display for review. Therefore,
according to this means, all generated particle images are not
always stored in the review image memory, so that the memory
capacity can be minimized and since there is no need to reduce the
number of review particles to be stored originally regardless of
it, the reproducibility of the classification result can be
prevented from reduction.
The solution means of 2 displays particle images in batch for each
biological particle kind. Therefore, according to this means, the
review work can be simplified.
The solution means of 3 designates the particle kind for one sample
in each measuring mode. Therefore, according to this means, useless
particle images can be prevented from storage in the review image
memory.
The solution means of 4 comprises a means for designating the
particle kind to be reviewed, a review image memory, a means for
removing images of particles corresponding to particle kinds other
than the particle kind to be reviewed and controlling storage of an
image in the review image memory so as to store an image of
particles corresponding to the designated particle kind in the
review image memory, and a means for displaying the stored particle
image for review. Therefore, according to this means, in the same
as with the solution means of 1, all particle images are not always
stored in the review image memory, so that the memory capacity can
be minimized and since there is no need to reduce the number of
review particles to be stored originally regardless of it, the
reproducibility of the classification result can be prevented from
reduction.
In the solution means of 5, the image storage control means
includes a means for turning storage of an image of particles which
cannot be classified on or off and is structured so as to store the
image of particles which cannot be classified in the review image
memory when the image storage ON/OFF means is in the ON state.
Therefore, according to this means, particles which cannot be
classified by the apparatus can be classified by an operator.
The solution means of 6 has a means for registering a specific
particle kind beforehand and the image storage control means
includes a means for turning storage of an image of particles
corresponding to the specific particle kind on or off and is
structured so as to store an image of particles which correspond to
the registered specific particle kind and are automatically
classified in the review image memory when the image storage ON/OFF
means is in the ON state. Therefore, according to this means, when,
for example, particularly particles which are medically important
are selected as a specific particle, these particles can be
reviewed by an operator even if they are not designated one by
one.
In the solution means of 7, the image storage control means
includes a means for turning storage of an image of ambiguous
particles which are considered to correspond to the particle kind
of the first candidate but are considered to possibly correspond to
the particle kind of the second candidate close to the particle
kind of the first candidate on or off and is structured so as to
store the image of ambiguous particles in the review image memory
when the image storage ON/OFF means is in the ON state and the
display means is structured so as to display the particle kinds of
both the candidates in addition to the image of ambiguous particles
which is stored as an image relating to the particle kinds of both
the candidates. Therefore, according to this means, ambiguous
particles which cannot be classified definitely by the apparatus
can be classified efficiently by review by an operator.
In the solution means of 8, the image storage control means
includes a means for turning storage of an image of particles on or
off and is structured so as to store an image of particles
corresponding to another particle kind which is apt to be
misdecided as one corresponding to the designated particle kind by
automatic classification in the review image memory when the image
storage ON/OFF means is in the ON state and the display means is
structured so as to display the image of particles corresponding to
the designated particle kind and the image of particles
corresponding to the another particle kind for comparison
observation. Therefore, according to this means, particles which
cannot be classified correctly by the apparatus can be classified
efficiently by an operator.
The solution means of 9 is structured so that an operator
classifies particles displayed on the display unit finally or
changes the particle kind name and furthermore corrects the
automatic classification result of particles which are
automatically classified on the basis of the aforementioned
classification result. Therefore, according to this means, the
particle classification and discrimination precision can be
improved.
The solution means of 10 is structured so that the automatic
classification result of the automatically classified particles is
corrected on the basis of the existence ratio of the particles
which are objects of review for each corresponding particle kind.
Therefore, according to this means, even if the capacity of the
review image memory is insufficient, the particle classification
and discrimination precision can be prevented from reduction due to
it.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a whole block diagram of a flow type particle image
analytical apparatus which is an embodiment of the present
invention.
FIG. 2 is a perspective view showing the constitution of the flow
cell shown in FIG. 1.
FIG. 3 is a flow chart for the review work as an example used
together with the apparatus shown in FIG. 1.
FIG. 4 shows a display screen of review images of the display unit
shown in FIG. 1 as an example.
FIG. 5 is a flow chart for the review work as another example used
together with the apparatus shown in FIG. 1.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
FIG. 1 is a whole schematic block diagram of a flow type particle
image analytical apparatus which is an embodiment of the present
invention. The flow type particle image analytical apparatus
comprises a flow cell 100, an image pick-up device 101, a particle
analyzer 102, a particle detector 103, and a flow system controller
124.
The image picking-up device 101 comprises a flash lamp driving
circuit 1a, a flash lamp 1, a field lens 2, a view field stop 11,
an aperture stop 12, a microscope condenser lens 3, a microscope
object lens 5 (shared by the particle detector 103), and a TV
camera 8. The particle analyzer 102 comprises an A-D converter 24,
an image memory 25, an image processing control circuit 26, a
characteristic extraction circuit 27, a discrimination circuit 28,
a particle number analyzer 40, a central control unit 29, a review
particle image memory 30, a display unit 50, and a key input unit
60.
The flow cell 100 comprises, as shown in FIG. 2, a parallel flow
path 150, a contraction flow path 151, a measurement flow path 152,
and a deceleration flow path 153. The flow cell 100 is generally
made of glass.
The cross section of the parallel flow path 150 perpendicular to a
sample flow direction 111 between an inlet 117 and the junction
portion with the contraction flow path 151 is quadrangular and a
nozzle 114 is installed in the neighborhood of the inlet 117 of the
parallel flow path 150.
The nozzle 114 has a cross section with a rectangular shape in
which the thickness direction thereof which is almost the same as
the passing direction of flashlight which will be described later
is a short side and the width direction which is perpendicular to
the thickness direction and the sample flow direction is a long
side. The intersection point of the diagonal lines of the rectangle
coincides with the intersection point of the diagonal lines of the
quadrangle of the cross section of the inlet 117 of the parallel
flow path 150 and the rectangle is contained inside the quadrangle.
The inside of the nozzle 114 is a flow path of a sample and the
outside thereof is a flow path of a sheath solution. Numeral 112
indicates a flow direction of a sheath solution.
The cross section of the nozzle 114 up to a nozzle outlet 116 in
the sample flow direction 111 has a rectangular shape which is
almost the same. A sample guide 113 is attached to the nozzle
outlet 116 so as to keep the width of sample flow constant. The
sample guide 113 comprises a pair of plates which face each other
across the sample flow and is extended near the center of the
parallel flow path 150 from the nozzle outlet 116.
The cross section of the contraction flow path 151 between the
junction portion with the parallel flow path 150 and the junction
portion with the measurement flow path 152 is quadrangular so that
the size in the width direction is kept unchanged and the size in
the thickness direction reduces slowly toward the measurement flow
path 152.
The cross section of the measurement flow path 152 between the
junction portion with the contraction flow path 151 and the
junction portion with the deceleration flow path 153 has the same
quadrangular shape and a particle detection area 80 and an image
picking-up area 90 are installed at the center thereof.
The particle detection area 80 is extended in the width direction
mentioned above and has a long and narrow shape having a length
which is the same as the sample flow width. The image picking-up
area 90 is arranged on the downstream side of the particle
detection area 80 and has a quadrangular shape, one side of which
has a length which is almost the same as the sample flow width.
The cross section of the deceleration flow path 153 between the
junction portion with the measurement flow path 152 and an outlet
118 is quadrangular, and the size in the width direction is
constant, and the size in the thickness direction enlarges slowly
along the sample flow direction.
Next, the status of flow of a particle suspended sample and a
sheath solution inside the flow cell 100 will be explained.
The sample containing suspended particles 160 and the sheath
solution flow into the parallel path 150 from the inlet 117. In the
parallel flow path 150, the sample and sheath solution flow in
according to the shape of the nozzle 114, so that a double fluid
layer having an inner layer of the sample and an outer layer of the
sheath solution (coated layer) is formed.
The guide 113 of the nozzle 114 suppresses the disturbance of the
liquid sample at the nozzle outlet 116. The width of the sample can
be kept at almost the width of the guide 113 between the nozzle
outlet 116 and the outlet 118. When the ratio of flow rate between
the sample and the sheath solution is changed by the guide 113, the
width of the sample is kept constant and only the thickness
changes.
When a liquid flows into the contraction flow path 151, the liquid
is contracted only in the thickness direction and a superflat
sample flow with a width of 200 to 300 m and a thickness of several
to tens m is formed.
When the superflat sample flow passes through the measurement flow
path 152, the particles 160 in the sample are detected in the
particle detection area 80 and imaged in the image picking-up area
90. Then, the superflat sample flow passes through the deceleration
flow path 153 and reaches the outlet 118.
In the flow cell 100, the thickness of the superflat sample flow in
the measurement flow path 152 is adjusted according to the ratio of
flow rate between the sample and the sheath solution. For example,
if the flow rate of the sheath solution decreases when the flow
rate of the sample is kept constant, the width of the superflat
sample flow is kept constant and the thickness thereof increases.
If the flow rate of the sheath solution increases, the width of the
superflat sample flow is kept constant and the thickness thereof
decreases. The particle detector 103 comprises a semiconductor
laser source 15, a collimator lens 16, a cylindrical lens 17, a
reflecting mirror 18, a minute reflecting mirror 19, a microscope
object lens 5, a beam splitter 20, a stop 21, a light detection
circuit 22, and a flash lamp lighting control circuit 23. A laser
beam from the semiconductor laser source 15 is changed to a
parallel laser flux 14 by the collimator lens 16 and the laser flux
14 is focused only in one direction by the cylindrical lens 17.
This focused laser flux is irradiated to the particle detection
area 80 in the flow cell 100 via the reflecting mirror 18 by the
minute reflecting mirror 19 arranged between the microscope lens 3
and the flow cell 100.
The particle detector 103 detects particles according to the
particle judgment logic for judging the presence or absence of
particles. There are a plurality of particle judgment logics
provided. When particles with small diameters are detected, a
detection signal from the light detection circuit 22 is on the A
level and the judgment logic (algorithm) for judging that particles
are detected when the pulse width becomes PA is used. When
particles with large diameters are detected, a detection signal
from the light detection circuit 22 is on the B level which is
different from the A level when small particles are detected and it
is judged that particles are detected when the pulse width becomes
PB which is different from PA.
In addition to the aforementioned judgment logic, it is possible to
use a judgment logic for judging that particles are detected from a
change in the waveform of a detection signal from the light
detection circuit 22 and to change the judgment level according to
the magnitude of the diameter of particles to be measured. When a
staining solution is added to a sample solution, it is possible to
detect particles by the color level and change the judgment color
level according to the magnitude of diameter of particles to be
measured.
The particle analyzer 102 converts picture element data signal of
an image outputted from the TV camera 8 to a digital signal by the
AD converter 24 and stores the data based on it at the
predetermined address of the image memory 25 under control of the
image processing control circuit 26. The data stored in the image
memory 25 is read under control of the image processing control
circuit 26 and inputted to the characteristic extraction circuit 27
and the discrimination circuit 28 so as to be image-process ed and
the result is supplied to the central control unit 29. The particle
classification result and the particle discrimination
characteristic parameter data used for particle classification are
supplied. The particle classification and discrimination process is
automatically executed by the pattern discrimination process which
is generally executed based on configurational features relating to
configuration, size or color of the particle image. This image
processing result, measurement conditions, and processed-image
information are sent from the central control unit 29 to the
particle number analyzer 40. The particle number analyzer 40 checks
the correspondence between the detected particles and the particle
classification result on the basis of a particle detection signal
from the central control unit 29 and the light detection circuit 22
and a control signal from the image processing control circuit 26
and summarizes the final classification and discrimination result
of particle images. The result is returned to the central control
unit 29 and outputted and displayed on the display unit 50 when
necessary.
On the other hand, to review a particle image, the particle name to
be reviewed is inputted by an operator from the key input unit 60
first and transmitted to the discrimination circuit 28 via the
central control unit 29. Only when the result classified and
discriminated according to biological kind of the particles by the
discrimination circuit 28 matches with the set review particle
name, is the corresponding particle image sent to the review image
memory 30 from the image memory 25 and stored sequentially. As to a
particle image, when even a dedicated particle image to be reviewed
is a particle image stored in the review image memory 30, it is
displayed on the display screen of the display unit 50 for each
same particle kind from the review image memory 30 via the central
control unit 29 after the sample measurement ends and used for
review by the operator.
The particle concentration in the sample and the view field
converted particle number are calculated on the basis of these
measurement results and the analytical result is returned to the
central control unit 29.
A flow system control unit 124 adjusts the ratio of flow rate
between a sample and a sheath solution introduced into the flow
cell 100 by a signal from the central control unit 29.
Next, the operation of a flow type particle image analytical
apparatus of an embodiment of the present invention will be
explained.
In FIG. 1, the sample and sheath solution flow in the flow cell 100
from the top of the drawing to the bottom thereof. The laser beam
from the semiconductor laser source 15 passes through the
collimator lens 16 and is changed to the laser flux 14. The laser
flux 14 is irradiated to the flow cell 100 via the cylindrical lens
17 and the reflecting plates 18 and 19. The laser beam passing
through the flow cell 100 is reflected from the beam splitter 20
via the microscope object lens 5 and irradiated to the light
detection circuit 22 via the stop 21.
When particles in the sample reach the laser beam passing position
in the flow cell 100, a detection signal is sent to the particle
number analyzer 40 and the flash lamp lighting control circuit 23
from the light detection circuit 22. By this detection signal, the
flash lamp lighting control circuit 23 lets the flash lamp 1 light
via the flash lamp driving circuit 1a. The flash light from the
flash lamp 1 passes through the lens 2 and is irradiated to the
particles in the flow cell 100 via the view field stop 11, the
aperture stop 12, and the microscope condenser lens 3. An image of
the irradiated particles is sent to the TV camera 8 via the
microscope object lens 5 so as to be picked up. The image
processing control circuit 26 supplies a command signal to the
image memory 25 and the particle number analyzer 40 according to
this information from the TV camera 8. The image information
obtained by the TV camera 8 is supplied to the image memory 25 via
the A-D converter 24. The image information is sent to the central
control unit 29 from the image memory 25 via the characteristic
extraction circuit 27 and the discrimination circuit 28.
The central control unit 29 controls the operations of the particle
number analyzer 40, the image processing control circuit 26, and
the flow system control unit 124 and lets the display unit 50
display a processed particle image.
To review a particle image, the name of a particle kind to be
reviewed is inputted by an operator from the key input unit 60
first before starting sample measurement and transmitted to the
discrimination circuit 28 via the central control unit 29. Only
when the result classified and discriminated by the discrimination
circuit 28 matches with the preset name of review particle kind, is
the corresponding particle image sent to the review image memory 30
from the image memory 25 and stored sequentially. The stored
particle image is fetched from the review image memory 30 after the
sample measurement ends and displayed on the display screen of the
display unit 50 for each same particle kind via the central control
unit 29. As a review operation for particles, an operator
classifies particles finally or changes the classification name by
observing a displayed particle image. The review result is
transmitted to the central control unit 29 and summarized as a
final particle classification result.
According to this embodiment, only images of particles
corresponding to the designated particle kind among particles which
are automatically classified are stored in the review image memory
30 but all the imaged particles are not stored in the review image
memory. Therefore, the review image memory 30 does not require a
large memory capacity.
Particle images to be reviewed are displayed on the CRT display
screen in batch for each particle kind and hence the review work is
simplified.
A large amount of particles of the same kind may be contained in a
sample. Under the condition that the automatic particle
classification precision of those particles is high, there is no
need to designate as review particles for those particles. Under
this particle analytical condition, if only particles of kinds
other than the above are designated as review particles, remarkable
saving of the capacity of the review image memory can be
expected.
To classify biological cells, by designating only particles which
are important particularly for medical diagnosis as review
particles, accomplishment of the measurement object can be
expected. The number of such specific particles is small and by
designating only specific cells as review particles, the review
time can be shortened and the work can be executed simply.
It is possible to switch the measurement mode for one sample and
designate the particle kind as a review particle in each switched
measurement mode. By doing this, a useless particle image will not
be stored in the review image memory. As an actual example of the
measurement mode, a case that a sample is measured under the
condition of a different flow rate, a case that a sample is
measured under the condition of a different image magnification, or
a case that a sample is measured under a different particle
detection condition (the magnitude of the detection signal level,
the pulse width of the signal, etc.) may be cited.
An image of a particle corresponding to the particle kind which
cannot be classified by the image process by the apparatus may be
stored in the review image memory 30. In this case, such a particle
which cannot be classified can be classified by the particle image
review function by watching the image by an operator and
improvement of the classification and discrimination precision can
be expected.
A review flow chart of a particle image which cannot be classified
as mentioned above is shown in FIG. 3. With reference to the
drawing, whether the normal process is switched to the storing
process of an image of a particle which cannot be classified or
not, that is, the storing process of the latter is turned on or off
is designated first (S1). In the drawing, it is described simply as
"process switching on/off". Next, the particle kind to be reviewed
is designated (S2), and then a particle is detected as usual, and
an image of the detected particle is inputted (S3) and processed so
as to extract a characteristic parameter (S4), and also the
particle classification and discrimination process is executed by
the pattern recognition process (S5). Thereafter, whether the image
is an image of a particle corresponding to the particle kind
designated to be reviewed or not is judged (S6). When the result is
yes, the particle image and classification result are stored in the
review image memory (S7). Even if the image is not an image of a
particle corresponding to the designated particle kind, whether the
image storage at Step S1 is in the ON state or not is judged (S8).
Furthermore, only when the answer is yes, whether the corresponding
particle image is an image of a particle which cannot be classified
or not is judged (S9). When the answer is yes, the image and
classification result are stored in the review image memory. By
doing this, it is found that when a particle cannot be classified,
the image thereof and classification result are stored in the
review image memory regardless of whether the particle is a
particle corresponding to the particle kind designated to be
reviewed.
Thereafter, the calculation for each classification is executed
(S10) and whether the processing is completed or not is judged
(S11). When the processing ends, the review work for the review
particle is executed. Namely, whether the corresponding particle is
a review particle or not is judged (S12). When the answer is yes,
the image of the particle corresponding to the designated particle
kind is displayed and the review work is executed by observing the
image (S13). Needless to say, data is corrected if necessary. Then,
undesignated review particles are also displayed and reviewed and
data is corrected if necessary (S14). Finally, whether the review
work ends or not is judged (S15). When the review work ends, the
final classification result which is reflected by the review result
is outputted (S16).
As to an image of an ambiguous particle which is considered to
correspond to the particle kind of the first candidate but is
considered to possibly correspond to the particle kind of the
second candidate close to the particle kind of the first candidate,
it is possible to store it in the review image memory. This
particle image can be reviewed almost in the same way as that shown
in FIG. 3. The reason is that if Step S1 is considered as a step of
designating whether the storage of an image of an ambiguous
particle is turned on or off and Step S9 is considered as a step of
judging whether the corresponding particle is close to the particle
kind of the second candidate or not, the flow shown in FIG. 3 can
be used as a flow for reviewing an ambiguous particle. Therefore,
the explanation of review of an ambiguous particle by referring to
FIG. 3 will be omitted so as to avoid duplication.
For an ambiguous particle, it is desirable to display the particle
kind names of the two candidates in addition to the image of the
particle from a viewpoint of improving the efficiency of the
classification work.
As to a particle corresponding to another particle kind which is
apt to be misdecided as one corresponding to the designated
particle kind by automatic classification, it is possible to store
it in the review image memory. Also in this case, the particle can
be reviewed almost in the same way as that shown in FIG. 3. The
reason is that if Step S1 shown in FIG. 3 is considered as a step
of designating whether the storage of an image of a particle
corresponding to another particle kind which is apt to be
misdecided as one corresponding to the designated particle kind is
turned on or off and Step S9 is considered as a step of judging
whether the corresponding particle is a particle which is apt to be
misdecided or not, the flow shown in FIG. 3 can be substituted for
a flow for reviewing the particle. Therefore, the explanation of
review of the particle by referring to FIG. 3 will be omitted so as
to avoid duplication.
For particles which are apt to be misdecided as mentioned above, it
is desirable to display an image of a particle corresponding to the
particle kind which is originally designated and a fixed number of
images of particles corresponding to another particle kind which is
apt to be misdecided at positions close to each other so as to
allow mutual comparison observation. FIG. 4 shows a display example
of an actual review image of the display unit 50 in that case.
In FIG. 4, images of particles corresponding to the particle kind
designated as a review particle are displayed in the particle
display area A and images of particles corresponding to a particle
kind which is apt to be misdecided are displayed in the particle
display area B. By doing this, an operator can proceed with the
classification work of the particles efficiently by comparing and
observing the images displayed in both the display areas. Needless
to say, when he classifies particles finally through the
classification review work or changes the particle kind name, the
automatic classification result which is reflected and corrected by
the result is outputted. Therefore, by doing this, the particle
classification and discrimination precision can be improved.
To classify biological cells, particularly as to a particle kind
which is important for medical diagnosis, it is significant to
decide it as a specific particle kind, register it in the memory
inside the central control unit 29, and store it in the review
image memory when the storage is in the ON state. The reason is
that there is no need to designate each time.
Also in this case, particles can be reviewed almost in the same way
as that shown in FIG. 3. The reason is that if Step S1 shown in
FIG. 3 is considered as a step of designating whether the storage
of an image of a particle corresponding to the specific particle
kind is turned on or off and Step S9 is considered as a step of
judging whether the corresponding particle is the specific particle
or not, the flow shown in FIG. 3 can be substituted for a flow for
reviewing the particle. Therefore, the explanation of review of the
particle by referring to FIG. 3 will be omitted so as to avoid
duplication.
When there are many particles stored in the review image memory and
the capacity of the review image memory is not sufficient, it is
desirable to correct the automatic classification result by the
particle existence ratio of reviewed particle images. The reason is
that by doing this, even if the memory capacity becomes
insufficient, the classification and discrimination precision will
not be lowered.
The flow of this work will be explained by referring to FIG. 5.
When the particle image analysis starts, the review particle is
designated first (S1), and then the particle detection and image
input (S2), the characteristic parameter extraction by image
processing (S3), and the particle classification and discrimination
process by pattern recognition (S4) are executed as in the above
explanation. When it is judged next at Step S5 that the review
particle is designated and it is furthermore judged at Step S6 that
the review image memory is not over the capacity thereof, the image
of the particle is stored in the review image memory (S7), and the
particle kind name thereof is also stored (S8) in it, and the
calculation for each classification is executed (S9).
However, the capacity of the review image memory is limited and
there may be a case that images of all review candidate particles
cannot be stored fully. If this occurs, "memory over" is set (S10)
and the storage of images into the review image memory is stopped.
The aforementioned series of operations is continued for one sample
and when the processing ends (S11), the review operation is
started.
As to the review operation, an image of the review-designated
particle is displayed on the display unit and the classification
result is classified again or finally by an operator (S12). Next,
when "memory over" is set (S13), the existence (classification)
ratio is calculated for each classification of a particle
corresponding to the review-designated particle kind (S14).
Finally, the whole particle classification result is calculated in
consideration of the classification ratio of review particles
obtained by the review operation for all the designated particles
(S15) and the result is outputted (S16).
As a particle detection means, a case where a laser flux from a
semiconductor laser is used as detection light and the laser flux
scattered by particles is used, is described. However, there is no
limit to it. Fluorescence from particles or transmitted light may
be used, or a method for detecting particles by a one-dimensional
image sensor or a method for detecting particles by a change in the
electric resistance due to passage of particles may be used.
In the embodiment, a flow cell keeping the width of a sample
constant is used. However, the present invention can be applied to
an apparatus using a flow cell for enlarging or reducing the width
of a sample slowly in the imaging area.
Effects of the present invention are as follows:
1. The present invention designates the kind of particles to be
reviewed, controls storage of an image in a review image memory so
as to store an image of particles corresponding to the designated
particle kind in the review image memory, and displays the stored
particle image on the display for review, so that all generated
particle images are not always stored in the review image memory.
Accordingly the memory capacity can be minimized and since there is
no need to reduce the number of review particles to be stored
originally regardless of it, the reproducibility of the
classification result can be prevented from reduction.
2. The present invention displays particle images in batch for each
particle kind, so that the review work can be simplified.
3. The present invention designates the particle kind for one
sample in each measuring mode, so that useless particle images can
be prevented from storage in the review image memory.
4. The image storage control means includes a means for turning
storage of an image of particles which cannot be classified on or
off and is structured so as to store the image of particles which
cannot be classified in the review image memory when the image
storage ON/OFF means is in the ON state, so that particles which
cannot be classified by the apparatus can be classified by an
operator.
5. The present invention has a means for registering a specific
particle kind beforehand and the image storage control means
includes a means for turning storage of an image of particles
corresponding to the specific particle kind on or off and is
structured so as to store an image of particles which correspond to
the registered specific particle kind and are automatically
classified in the review image memory when the image storage ON/OFF
means is in the ON state, so that when, for example, particularly
particles which are medically important are selected as a specific
particle, these particles can be reviewed by an operator even if
they are not designated one by one.
6. The image storage control means includes a means for turning
storage of an image of ambiguous particles which are considered to
correspond to the particle kind of the first candidate but are
considered to possibly correspond to the particle kind of the
second candidate close to the particle kind of the first candidate,
on or off and is structured so as to store the image of ambiguous
particles in the review image memory when the image storage ON/OFF
means is in the ON state and the display means is structured so as
to display the particle kinds of both the candidates in addition to
the image of ambiguous particles which is stored as an image
relating to the particle kinds of both the candidates, so that
ambiguous particles which cannot be classified definitely by the
apparatus can be classified efficiently by review by an
operator.
7. The image storage control means includes a means for turning
storage of an image of particles on or off and is structured so as
to store an image of particles corresponding to another particle
kind which is apt to be misdecided as one corresponding to the
designated particle kind by automatic classification in the review
image memory when the image storage ON/OFF means is in the ON state
and the display means is structured so as to display the image of
particles corresponding to the designated particle kind and the
image of particles corresponding to the another particle kind for
comparison observation, so that particles which cannot be
classified correctly by the apparatus can be classified efficiently
by an operator.
8. The present invention is structured so that an operator
classifies particles displayed on the display unit finally or
changes the particle kind name and furthermore corrects the
automatic classification result of particles which are
automatically classified on the basis of the aforementioned
classification result, so that the particle classification and
discrimination precision can be improved.
9. The present invention is structured so that the automatic
classification result of the automatically classified particles is
corrected on the basis of the existence ratio of the particles
which are objects of review for each corresponding particle kind,
so that even if the capacity of the review image memory is
insufficient, the particle classification and discrimination
precision can be prevented from reduction due to it.
* * * * *